Faster Horses: The Trap of Chasing the Newest AI Model

There’s a line usually credited to Henry Ford: “If I had asked people what they wanted, they would have said faster horses.” Ford didn’t tweak the old paradigm. He changed it. Most businesses buying AI right now are asking for faster horses — a newer model, a shinier tool — when the thing that would actually help them is a different question entirely.

Chasing “newer” is optimizing the wrong axis

I watched an entire industry make this mistake. In shrimp farming, everyone chased faster-growing genetics to “outrun” disease. All that speed bred fragility — the crops failed harder, costs climbed, and the same people chasing speed were the loudest complainers about crop failures. Self-inflicted.

AI has its own version. A project underperforms, so the team reaches for the newest model, sure that the next version will fix it. It almost never does. The jump from last year’s model to this year’s rarely changes your ROI — the workflow around it does. You’re optimizing horsepower when the problem is the wagon.

The model you already have is almost certainly good enough

For the overwhelming majority of real business problems, the constraint isn’t model quality. It’s a fuzzy problem definition, a broken process, or data nobody trusts. Point a merely-good model at a sharply-defined problem and you’ll beat a state-of-the-art model aimed at a vague one every time.

So before you upgrade anything: name the specific problem, in dollars or hours. Nine times out of ten, the answer isn’t a faster horse. It’s aiming what you already have at the right target.